Re: [gfortran, PR33254] Add -fbounds-checking for character array constructors

[Tobias Schlüter <tobias dot schlueter at physik dot uni-muenchen dot de>]

Dominique Dhumieres wrote:
> catches the error, but if I put y(1:len(trim(y))) as the first item,
> it does not:
>
> [karma] f90/bug% diff bounds_check_10_db_1.f90 bounds_check_10_db.f90
> 12c12
> < z = [trim(x), y(1:len(trim(y))), y(1:len(trim(x)))] ! should catch first error
> ---
> > z = [y(1:len(trim(y))), trim(x), y(1:len(trim(x)))] ! should catch first error
> [karma] f90/bug% gfc -fbounds-check bounds_check_10_db.f90
> [karma] f90/bug% a.out
> [karma] f90/bug% 
>
> Am I alone to see that? If not, any idea about the reason?

It's really weird.  In the case where it doesn't work the bounds-checks 
simply don't appear in the .original dump.

I'll investigate.

Cheers,
- Tobi